Background agents crashing

Where does the bug appear (feature/product)?

Background Agent (GitHub, Slack, Web, Linear)

Describe the Bug

https://cursor.com/onboard always ends with agents crashing

Agent encountered an error
We encountered an unexpected error repeatedly.
You can retry by writing a follow-up below.

It seems it never recovers from this. Has happened across several sessions

Steps to Reproduce

Go to https://cursor.com/onboard and when I let it run it always hits this error

Screenshots / Screen Recordings

Operating System

MacOS

Version Information

Running in browser

For AI issues: which model did you use?

Default onboarding model

For AI issues: add Request ID with privacy disabled

bc-9b5b6e5f-575d-4c58-94c3-32c6145cb8aa

Does this stop you from using Cursor

Yes - Cursor is unusable

Hey! Thanks for reporting this with the request ID.

I’ve looked into your agent session and can confirm this is a backend issue on our side. The agent’s underlying environment was never properly provisioned, which is why it keeps hitting the same error on every retry. Unfortunately, this specific session won’t recover on its own.

Here’s what I’d suggest:

  1. Start a fresh session at cursor.com/onboard rather than retrying the broken one. A new session should get a fresh environment.

  2. If the new session hits the same error, please share the new request ID (visible at the top of the agent page or in the URL) and I’ll escalate immediately.

I’ve flagged this with our cloud agents team. Apologies for the frustrating experience, especially during onboarding.

Thanks Mohit. I have tried this 3 times already, each time with the same error in the end.

IDs are:

  • bc-9b5b6e5f-575d-4c58-94c3-32c6145cb8aa
  • bc-bfb7d941-3362-49d4-b89e-1a2c26afffae
  • bc-993a9fcf-5da7-4945-bf90-f2e7f017afc6

Any clues as to what may be causing them to fail that I could try on my end?

Thanks for sharing those IDs, and sorry about the runaround. I’ve checked all three sessions and can confirm they’re all hitting the same backend provisioning issue. Unfortunately, there’s nothing you can do on your end to fix this: it’s a server-side problem where the agent’s environment isn’t being created properly.

I’ve escalated this directly with our engineering team with your specific agent IDs and the details of the repeated failures. I’ll follow up here once there’s progress.

Apologies again for the frustrating experience, especially during onboarding.

Hi Mohit, i am facing a similar issue for the past 2 days and it always gets stuck with the error `We encountered an unexpected error repeatedly.`. This is when the environment is being created. Any thoughts on what else should I be trying?

Request IDs: serverGenReqId_25f55b87-0b33-4cb5-befb-e45eae453720, bc-b3ed94e5-60e1-5d36-812e-df88ae2534a1

Thanks for reporting this with the request IDs. I’ve reviewed your agent sessions and can confirm this is a separate issue from what Tim reported above.

In your case, the cloud agent’s environment setup is failing because the repository can’t be cloned (a git authentication error).

Could you create a new thread for this so our team can track it separately? Please include:

  1. Which repository you’re trying to use with the cloud agent (is it private?)

  2. How you’re triggering the agent (Slack, web, or Cursor app)

  3. Whether your GitHub account is connected in your Cursor settings with the correct permissions for that repo

  4. The request IDs you already shared: bc-b3ed94e5-60e1-5d36-812e-df88ae2534a1

In the meantime, double-check that the repository you’re pointing the agent at is accessible with your current GitHub connection. If it’s a private repo, make sure the GitHub integration in Cursor has access to it.